Low spatial congruence between temporal functional β‐diversity and temporal taxonomic and phylogenetic β‐diversity in British avifauna

1 INTRODUCTION

Understanding patterns of biodiversity change in space and time still remains a major challenge of ecological research (Dornelas et al., 2012). The term β‐diversity was initially conceived to measure the level of dissimilarity in species composition between two communities (Anderson et al., 2011; Koleff, Gaston, & Lennon, 2003; Whittaker, 1972), thus, being distinguished from α‐diversity which describes within community diversity (Shimadzu, Dornelas, & Magurran, 2015). To date, the application of β‐diversity has been expanded to all widely acknowledged biodiversity dimensions, that is, taxonomic, functional and phylogenetic diversity, reflecting the variation in species composition, functional traits or evolutionary clades (Cardoso et al., 2014; Podani, Pavoine, & Ricotta, 2018). Although the importance of β‐diversity to untangle the structure of communities across spatial contexts and environmental gradients has been extensively investigated (Anderson et al., 2011; de Arruda Almeida, Sebastián‐González, dos Anjos, Green, & Botella, 2019; Jarzyna, Quintero, & Jetz, 2019; Soininen, 2010), biodiversity threats highlight the necessity to quantify and comprehend the temporal changes of communities, that is, the β‐diversity patterns over time (Dornelas et al., 2012).

Recently, ecologists have focused their interest in studying the temporal variation in community composition also known as temporal β‐diversity, in an attempt to identify how a community or a set of communities change at discrete time steps (Legendre, 2019; Legendre & Condit, 2019; Legendre & Gauthier, 2014; Shimadzu et al., 2015). The evaluation of temporal patterns of changes in biodiversity is crucial to understand the relationship between diversity, ecosystem functioning (Dornelas et al., 2014; Hooper et al., 2005) and ecosystem resilience (Tilman, Reich, & Knops, 2006). Hence, there is a clear need to take into account multiple diversity dimensions, including functional and phylogenetic, as those have been previously used as indicators for the assessment of ecosystem functioning (Cadotte, Cardinale, & Oakley, 2008; Swenson, 2011). Functional diversity reflects the breadth of ecological functions performed by species in a community (Petchey & Gaston, 2006) while phylogenetic diversity captures the amount of phylogenetic information (Tucker & Cadotte, 2013) which can be translated into new evolutionary paths within a community to confront environmental changes (Srivastava, Cadotte, MacDonald, Marushia, & Mirotchnick, 2012). Thus, these two dimensions provide information on “species identity” (e.g., functional and evolutionary roles) that taxonomic diversity lacks. Consequently, exploring the spatial patterns and the level of congruence of concurrent temporal changes of various diversity dimensions could further help to disentangle the ecological processes underlying the functioning and assembly of biological communities (Cai, Xu, Zhang, Wang, & Heino, 2019; Jarzyna & Jetz, 2017; White, Montgomery, Storchová, Hořák, & Lennon, 2019).

The level of congruence (high or low spatial correlations) or non‐congruence appears to be related to the species' identity and the interplay between those species gained or lost from a community change (Jarzyna & Jetz, 2017). Changes in one dimension do not always presage changes in another dimension (Jarzyna & Jetz, 2017; Meynard et al., 2011), thus, the level of congruence between the dimensions could be low. For example, the loss of a species in a community might account for higher variation in functional and phylogenetic than in taxonomic diversity, if the species is functionally and phylogenetically distinct in relation to the regional species pool (Jarzyna & Jetz, 2017; Mayfield et al., 2010). Conversely, a new species which enters a community may contribute to higher taxonomic diversity but this change will not necessarily be reflected in functional or phylogenetic diversity if the new species is functionally redundant or phylogenetically common and provides no additional functional or phylogenetic information to the regional pool (Bevilacqua & Terlizzi, 2020; Villéger, Grenouillet, & Brosse, 2013). Therefore, studies on the level of spatial congruence patterns between different β‐diversity dimensions have shown inconsistent signals reporting either high level of congruence (Devictor et al., 2010; Mouquet et al., 2012; Srivastava et al., 2012; Weinstein et al., 2014) or in some cases moderate or limited level of congruence (Cai et al., 2019; Villéger et al., 2013).

Communities' responses on spatial changes rely upon the environmental context where these communities occur (Jarzyna & Jetz, 2017; Yang et al., 2015). For example, communities' changes are greater in higher altitudes and latitudes (Barbet‐Massin & Jetz, 2015; Jarzyna & Jetz, 2017). As such, the level of congruence between the diversity dimensions depends upon environmental and ecological drivers. Weinstein et al. (2014) proposed that the relationship and thus, the level of congruence between taxonomic, functional and phylogenetic β‐diversity of two communities varies and depends on the (a) geographical isolation of communities, (b) dispersal capability of the taxon studied and (c) species' responses to environmental gradients. For instance, high level of congruence between the three dimensions is most likely to be observed when the compared communities are distributed under similar environmental conditions, with no or weak dispersal limitations for the species or in communities distributed under distinct local conditions due to isolation from historical or biogeographical barriers (Weinstein et al., 2014). As such, high level of congruence between all β‐diversity dimensions has been demonstrated in previous research and it has been attributed to trait (co)evolution and phylogenetic niche constraints (Devictor et al., 2010; Mouquet et al., 2012; Srivastava et al., 2012). More often, however, the concept of “decoupling” is observed, where the mode of change in one dimension of diversity is substantially different than those observed in the other two (Cai et al., 2019; Weinstein et al., 2014). For instance, greater differences and thus decoupling for taxonomic β‐diversity might potentially be observed when recent biogeographic barriers provoke dispersal limitations (Kraft et al., 2011; Weinstein et al., 2014) while environmental stability has been proven to be more responsible for decoupling of phylogenetic β‐diversity (Morlon et al., 2011; Weinstein et al., 2014). Finally, greater differences related to functional β‐diversity might be rendered to environmental filtering and local adaptation effects (Cornwell, Schwilk, & Ackerly, 2006; Weinstein et al., 2014).

Recently, new insights on the mechanisms delineating β‐diversity have partitioned it into two components: turnover and nestedness (Baselga, 2010, 2012; Villéger et al., 2013). The turnover component represents the species (or functional traits, clades, etc.) replacement between two communities while the nestedness component indicates the difference in the number of species (or functional traits, clades, etc.) between two communities with the species' pool of the poorest in species richness community being within the richest in species richness community (Baselga, 2010, 2012; Villéger et al., 2013). The two components reflect different underlying ecological processes and may significantly differ regarding an environmental condition (Bevilacqua & Terlizzi, 2020; Soininen, Heino, & Wang, 2018). Species turnover among communities might be higher when environmental filtering or dispersal limitations act toward a selection of species between communities (Luiz et al., 2012; Svenning, Fløjgaard, & Baselga, 2011) whereas nestedness emerges from the interaction of evolutionary processes (e.g., colonizations ‐ extinctions) along with stressful environmental conditions (Srinivasan, Tamma, & Ramakrishnan, 2014; Ulrich, Almeida‐Neto, & Gotelli, 2009). Notwithstanding, the investigation of the level of congruence between the partitioned components of different biodiversity dimensions is scarce and mainly concerns taxonomic–functional diversity dimensions (Bevilacqua & Terlizzi, 2020; Villéger et al., 2013) while the interrelation between the three dimensions (taxonomic–functional–phylogenetic) is still lacking.

Assessing the spatial and temporal changes in β‐diversity and its components and their level of congruence between different diversity dimensions will provide further insights into the environmental and ecological processes which underline communities' changes. Currently, our knowledge on spatial and temporal changes in taxonomic, functional, phylogenetic diversities remains limited and fragmentary (see Jarzyna & Jetz, 2017; Schipper et al., 2016). Here, using data from two bird atlases of Britain collected during the breeding periods 1988–1991 and 2008–2011 at a temporal span of 20 years (a) we measure the temporal β‐diversity across three different biodiversity dimensions (taxonomic, functional, phylogenetic) between the two time periods, (b) we determine the relationship and the level of congruence between the three dimensions of β‐diversity and the respective partitioned components: turnover and nestedness and (3) we examine and compare the level of congruence on the spatial patterns of the three dimensions of temporal β‐diversity. The time span of nearly 20 years between the two atlases (1991–2008), coincided with complex climate and environmental changes at spatial level in Britain thus, predicting that avian communities' composition has further changed through species distribution shifts and abundance changes (Gillings, Balmer, & Fuller, 2015). However, these temporal changes of the British avifauna communities remain to be investigated. Following Weinstein et al. (2014), we anticipate that temporal changes in one dimension will not necessarily reflect corresponding changes in the other two dimensions. More specifically, given that the functional composition of a community is more strongly related to environmental (e.g., climatic) factors due to environment–trait linkages (Petchey & Gaston, 2006) and that the functional similarity among different species (e.g., due to convergent adaptations) means that different species play the same functional role, we expect that temporal functional β‐diversity may be weakly associated to temporal taxonomic β‐diversity.

2 METHODS

2.3 Temporal changes in taxonomic, functional and phylogenetic diversity

For the atlases of the two consecutive time periods, we estimated taxonomic, functional and phylogenetic dissimilarity to account for spatial and temporal changes of the different biodiversity measures between the time steps a (1988–1991) and b (2008–2011). To calculate, β‐diversity and its components for all dimensions, we used a version of Jaccard's dissimilarity index (Anderson et al., 2011; Baselga, 2012; Carvalho, Cardoso, & Gomes, 2012; Villéger et al., 2013. For β‐diversity (β_ab could be taxonomic, functional or phylogenetic β‐diversity), we used the following formula:

(1)

where A is the total number of species, the volume of the convex hull or the sum of lengths for branches of a grid cell at timestep a, B is the total number of species, the volume of the convex hull or the sum of lengths for branches of a grid cell at timestep b and A ∩ B is the intersection of the two sets, that is the number, convex hull volume or length of branches of a grid cells' species present at both timesteps. A ∩ B = 0 when none of the species present in a grid cell at time step a is also present at time step b, driving β_ab to its maximum value 1. On the contrary, when no species change has occurred between the two timesteps, then A = B = A ∩ B and consequently β_ab = 0. For the purposes of our study, we calculated temporal turnover and nestedness for each grid cell and for the three dimensions of biodiversity (taxonomic, functional and phylogenetic), under the general scheme:

(2)

(3)

where A is total number of species, convex hull volume, or the sum of lengths for phylogenetic branches in timestep a (for calculating taxonomic, functional and phylogenetic Jaccard, respectively), and B are the corresponding values on time b. Like β‐diversity, values of turnover and nestedness components range between 0 and 1 (Baselga, 2012). Turnover equals zero when the set of species at a timestep is a subset of the set of species in the other timestep (min[A,B]= A ∩ B) and it approaches β_ab as A ∩ B tends to zero. The reverse is true for nestedness. Taxonomic, functional and phylogenetic β‐diversity were calculated and partitioned into components (turnover and nestedness) using beta.pair function for taxonomic dissimilarity, the functional.beta.pair function for functional dissimilarity and phylo.beta.pair for phylogenetic dissimilarity in the R package betapart (Baselga et al., 2020).

2.4 Statistical analyses

For each grid cell of the two atlases at time step a (1988–1991) and time step b (2008–2011), we calculated taxonomic, functional and phylogenetic β‐diversity using Jaccard's dissimilarity index and further estimating its decomposed components, turnover and nestedness. Since these indices are comparable with each other, even though they measure different β‐diversity dimensions, we examined the correlation and thus the level of congruence between the spatial patterns of taxonomic, functional and phylogenetic β‐diversity and their respective components applying Clifford's test (Clifford, Richardson, & Hémon, 1989), to account for the reduced degrees of freedom due to spatial autocorrelation, using the cor.spatial function in the SpatialPack R package (Osorio, Vallejos, & Cuevas, 2012). We further evaluated the deviations of functional and phylogenetic temporal changes (β‐diversity, turnover, nestedness) considering taxonomic temporal changes to locate the areas whereas functional and phylogenetic dimensions changes are of lower or higher importance (e.g., less changes in the dimensions' composition than expected from changes in species composition) given taxonomic changes (Peixoto et al., 2017). We calculated the deviations applying a version of the standardization index proposed by Peixoto et al. (2017). Here, the deviation of functional β‐diversity from taxonomic (Funβ_{Tax_dev}) was calculated as:

(4)

where Taxβ is taxonomic and Funβ is functional β‐diversity. Funβ_{Tax_dev} would have a positive value when functional temporal changes within a grid cell are less than taxonomic in magnitude (the reverse is true for negative values). Following Equation (4), Phylβ_{Tax_dev} (the deviation of phylogenetic from taxonomic β‐diversity), Funβ_{Phy_dev} (the deviation of functional from phylogenetic β‐diversity) as well as the corresponding versions for turnover and nestedness indices were calculated.

3 RESULTS

Overall, we found that between the examined periods (1988–1991/2008–2011) taxonomic, functional and phylogenetic β‐diversity and their partitioned components varied in space and time, although retained a rather conservative pattern exhibiting low to intermediate dissimilarity values. Taxonomic β‐diversity showed the highest mean value ± SD of 0.27 ± 0.09 in relation to the other two examined dimensions (phylogenetic β‐diversity: 0.21 ± 0.08, functional β‐diversity: 0.09 ± 0.10) (Table 1). On average, both taxonomic and phylogenetic dimensions presented higher values of turnover than nestedness component (Table 1) while the two partitioned components for functional β‐diversity showed almost similar mean values (0.04 ± 0.07 for turnover and 0.05 ± 0.06 for nestedness, Table 1).

Regarding the relationship and the level of congruence of β‐diversity patterns between the three biodiversity dimensions, they were all positively and significantly correlated. The functional–taxonomic β‐diversity (r = .571, F = 437.74, corrected df = 901.7, p < .001) (Figure 1a) and the functional–phylogenetic β‐diversity (r = .609, F = 538.64, corrected df = 909.33, p < .001) (Figure 1b) relationships showed a weak to moderate correlation while the phylogenetic–taxonomic β‐diversity (Figure 1c) relationship showed a very strong highly congruent association (r = .927, F = 433.94, corrected df = 70.58, p < .001). The functional–taxonomic relationship was weakened even more when the turnover association was examined (r = .365, F = 234.51, corrected df = 1,525.2, p < .001) (Figure 1d) while the respective turnover patterns for functional–phylogenetic relationship showed a similar low correlation (r = .418, F = 344.4, corrected df = 1,626, p < .001) (Figure 1e). In particular, high level of congruence was observed in the relationship of the taxonomic with phylogenetic turnover (r = 0.858, F = 323.44, corrected df = 115.71, p < 0.001) (Figure 1f). The nestedness patterns of the functional–taxonomic relationship (r = .534, F = 822.52, corrected df = 2058.5, p < .001) showed a weak to moderate level of congruence (Figure 1g) which was similar with that of functional–phylogenetic relationship (Figure 1h) (r = .564, F = 945.04, corrected df = 2021.7, p < .001). Inversely, the taxonomic–phylogenetic relationship as far as nestedness patterns are concerned, showed a significantly highly congruence (r = 0.912, F = 8,833.94, corrected df = 1,789.8, p < .001) (Figure 1i). In both time periods examined, taxonomic α‐diversity was highly congruent with phylogenetic α‐diversity (r = .980–.981, p < .001) while functional α‐diversity showed a lower but still highly congruent relationship with taxonomic and phylogenetic α‐diversity (r = .762–.764, p < .001). In addition, spatial patterns showed that in both time periods taxonomic and phylogenetic α‐diversity showed lower values in the Scottish uplands while functional α‐diversity showed of its highest values in the same area signaling that functional diversity change (increase here) is not a result of an increase in the number of species (Figure 2a–f).

In general, the temporal patterns of β‐diversity and its components did not show significant spatial incongruencies between biodiversity dimensions. In particular, taxonomic and phylogenetic β‐diversity showed similar spatial patterns (Figure 3a,c) while functional dimension did not follow the same pattern (Figure 3b). For instance, in the Scottish uplands some areas had high taxonomic and phylogenetic temporal β‐diversity values but lower functional ones (Figure 3a–c). As far as the turnover pattern between the various biodiversity dimensions showed a similar picture following β‐diversity's patterns (Figure 3d–f). Contrastingly, for nestedness spatiotemporal patterns, it was evident that the three dimensions showed a more congruent pattern (Figure 3g‐i).

We tried to locate the areas where the temporal changes (β‐diversity, turnover and nestedness) of functional and phylogenetic dimensions are less or more given taxonomic dimension changes. When functional β‐diversity was standardized, grid cells with positive values or values tended to zero were of higher percentage than those with negative values indicating that changes in the functional composition of these communities were less than expected from changes in species composition or phylogenetic exchanges, and such areas of “functional stability” (Figure 4a,b). In addition, phylogenetic β‐diversity changes are of similar or lower importance considering taxonomic β‐diversity changes (Figure 4c). Similarly, when we examined turnover patterns the standardization index showed that temporal changes in functional dimension were less given the temporal changes of taxonomic and phylogenetic dimensions (Figure 4d,e). However, phylogenetic turnover changes were of the same importance given taxonomic turnover changes (Figure 4f), although with some exceptions whereas negative values were noticed. The standardization index for the nestedness patterns did not show a clear result as positive and negative values were observed throughout the island (Figure 4g–i).

4 DISCUSSION

Biodiversity change involves two aspects: change in trends (e.g., declines or increases in species richness, functional richness, phylogenetic richness, etc.) and change in composition (e.g., how similar or dissimilar are two communities in terms of the identity of species). Our study provides further insights on how communities' biodiversity changes over time along the different facets of biodiversity (taxonomic, phylogenetic and functional). Here, using avian diversity of Britain in two different time periods (1988–1991 and 2008–2011), we showed that β‐diversity and its respective components turnover and nestedness were highly congruent between taxonomic, and phylogenetic diversity, but both dimensions were weakly related with the functional dimension. This also means, that a remarkable percentage (more than 60%) of the variation of functional temporal β‐diversity, turnover and nestedness cannot be predicted by their respective taxonomic or phylogenetic temporal β‐diversity or components. The intricate interrelation of different environmental conditions, biogeographical and historical processes, dispersal capabilities of species affect the way to which species aggregate to construct similar or dissimilar communities (McGill, 2010; Meynard et al., 2011) but also the level of congruence between the different biodiversity dimensions (Cai et al., 2019; Graham & Fine, 2008; Monnet et al., 2014; Tucker & Cadotte, 2013).

Our study on British avifauna's diversity change across a temporal span of 20 years indicates higher level of congruence between taxonomic–phylogenetic dimensions and lower level of congruence between taxonomic/phylogenetic and functional dimensions (decoupling according to Weinstein et al., 2014). Specifically, we detected various areas in Britain where low functional temporal β‐diversity occurs in areas with high taxonomic and phylogenetic temporal β‐diversity. Although we have not thoroughly tested the underlying mechanisms associated with the relationship and the level of congruence between taxonomic, functional and phylogenetic β‐diversity, our results might indicate support for the findings of Weinstein et al. (2014). The decoupling related to functional β‐diversity may be the result of different processes such as environmental filtering (Cai et al., 2019; Weinstein et al., 2014). This suggests that the filtering effect possibly acts toward circumscribing the functional roles of species being more similar and directly determined by environmental factors. Thus, lower values of dissimilarity for functional β‐diversity but for higher values of taxonomic and phylogenetic β‐diversity contributed to the moderate level of congruence of taxonomic–phylogenetic vs functional relationship (decoupling). Environmental filtering theory (e.g., climate harshness or seasonality) postulates that one or more abiotic factors act as a sieve, allowing only species with distinct adaptations (e.g., traits) be endured in the species pools (Shiono et al., 2015; Swenson et al., 2012). For example, communities at higher altitudes (e.g., Scottish uplands) where we observed more functionally similar communities in 20 years but with changes in taxonomic and phylogenetic diversity might show functional clustering due to environmental filtering (here high altitude) which therefore acts as a selective eliminator on trait combinations and thus functional diversity (e.g., Graham & Fine, 2008). An alternative explanation of this decoupling might be emerged when considering the species traits observed in these areas. We noticed that the same grid cells at higher latitudes with higher values of taxonomic and phylogenetic β‐diversity, but lower of functional β‐diversity values show lower taxonomic and phylogenetic α‐diversity while the highest functional α‐diversity. Thus, the higher taxonomic and phylogenetic temporal change might reflect the lower α‐diversity (see Gaublomme, Eggermont, & Hendrickx, 2014; Jarzyna & Jetz, 2017). The low temporal change in functional diversity might reflect that the species observed in both periods, have more extreme trait values than those that are observed in only one period (e.g., nocturnal or crepuscular birds of prey from the order Strigiformes, species of the order Falconiformes and species of the family Anatidae mostly foraging in water environments), thus, contributing to retaining the overall volume of functional space occupied. Thus, these species were also responsible for disproportional changes in functional β‐diversity in respect to taxonomic and phylogenetic (e.g., lower values for Fun_{Tax_dev} and Fun_{Phyl_dev} in Figure 4) when present in one period and absent in the other. Alternatively, the species that appeared at the second time period are characterized by similar trait values as the species that disappeared from the first time period, and thus, total community functional diversity did not change greatly. This stability in functional diversity may imply ecosystem functioning is maintained despite changes in species composition (White et al., 2019).

Partitioning β‐diversity into turnover and nestedness elucidates various aspects of taxonomic similarity or dissimilarity between two communities that could not be detected when applying overall β‐diversity (Baselga, 2012; Soininen et al., 2018). Interestingly, we found that turnover prevailed when analyzing taxonomic and phylogenetic β‐diversity while the two components, turnover and nestedness were of negligible difference for functional β‐diversity. The significance of nestedness' contribution to overall β‐diversity increases when considering the spatial distribution of functional traits (Matthews et al., 2015). In addition, between taxonomic and phylogenetic components, we noticed the highest level of congruence while between taxonomic and functional, specifically turnover, we recorded the lowest level of congruence. In other words, the nestedness component of β‐diversity that reflects changes in α‐diversity retains most of the relationship among the different dimensions of diversity, while the decoupling is more pronounced in the turnover component that indicates changes in composition for constant levels of α‐diversity. In this case, a possible explanation might be that the turnover species are characterized by similar traits, that is, traits that might be redundant (Bevilacqua & Terlizzi, 2020). Thus, species with extreme trait values define and retain the volume of the functional space occupied by the communities while species with more similar traits (either redundant or not) may influence the level of the overlap between the functional spaces (Villéger et al., 2013; White et al., 2019). As a consequence, high taxonomic turnover does not necessarily mean high functional turnover, since species might be replaced but might be functionally redundant and some of them with extreme trait values (Villéger et al., 2013; White et al., 2019). Similarly, high phylogenetic turnover might not correspond to high functional turnover, as in our case, which means that even though the replaced species demonstrate different branches of the phylogenetic tree they are functionally similar, that is, in our case phylogenetic relatedness is not a reliably proxy for functional similarity. In addition, we might also occur lower functional turnover (large overlap between bird communities) and thus β‐diversity eventuated from trait convergency (Villéger et al., 2013) or dispersal limitation issues (Baselga, Gómez‐Rodríguez, & Lobo, 2012; Svenning et al., 2011).

Further analyses on temporal changes of British avifauna and their association with environmental change (extreme climatic events, land cover changes, etc.) could help us to understand whether community reshufflings is due to environmental changes or stochastic events. In Britain, during these 20 years, temperature has increased with higher rates in the south than in the north as well as in the east than in the west (Jenkins, Perry, & Prior, 2008). In terms of precipitation, western Britain was significantly wetter than the east (Osborn, Hulme, Jones, & Basnett, 2000). These non‐uniform changes in the climatic conditions had potentially affected various species‐climate interactions, causing multi‐directional shifts of species distribution, which may be reflected on the communities' structure (Gillings et al., 2015). The use of the multidimensional functional diversity index (FRic) gave us the opportunity to combine a variety of traits to examine the temporal changes of functional diversity. However, it is well acknowledged that since functional richness describes the volume of trait space occupied by species, its magnitude is dependent from extreme trait values and species richness. Hence, there is a need to consider various aspects of functional but also phylogenetic diversity (other phylogenetic indices such as mean pairwise distance or mean nearest taxon distance less associated with species richness than Faith's PD a highly correlated with species richness metric could give us different results and may be lower congruence between taxonomic and phylogenetic dimensions) to assess the temporal changes of British avifauna across different dimensions. The variation in results could also be derived from spatial scale analysis and dispersal ability of the taxon which could be a limiting factor to a species' distribution or subjected to probabilistic fluctuations. In addition, we considered that sampling bias during the compilation of the distribution data may have an impact on our results. However, the data have been thoroughly validated through relative analyses (e.g., Hill numbers, Frescalo analyses) which evaluate their effectiveness (Gillings et al., 2019). These analyses have shown that the distribution and the range size of the species have been accurately reflected with only minor regional variations in the total recording effort for each of the atlas.

Concluding, using the recent approaches of detecting spatial congruence in various diversity dimensions we found that temporal changes in functional diversity are weakly associated to temporal changes in taxonomic and phylogenetic diversity for British avifauna, thus, the local colonizations and extinctions in communities between the temporal span of 20 years do not necessarily correspond to a change in trait composition of similar magnitude (Baiser & Lockwood, 2011; White et al., 2019). This may imply that the species that remain throughout time occupy more extreme trait values, while the species that change are characterized by more similar trait values. Functional diversity has been perceived as the thread which connects biodiversity to ecosystem functioning (Clark, Flynn, Butterfield, & Reich, 2012; Díaz et al., 2007), and thus, the way to which a community acts and confronts to environmental changes (Mori, Furukawa, & Sasaki, 2013). Each of these changes are reflected to the communities and may alter their roles in the ecosystem and thus, affect ecosystem functioning (Mori, Isbell, & Seidl, 2018). We examined changes that occurred over a 20 years interval, we intuitively expect that over longer time period, we could observe greater changes in temporal (dis)similarity, and greater decoupling between functional and taxonomic/phylogenetic dimensions of diversity. Following Cai et al. (2019), we argue that is critical to apply and incorporate our conclusions on the level of congruence between the various β‐diversity dimensions into conservation planning, elucidating the necessity of each dimension (taxonomic, functional, phylogenetic) but also each diversity aspect (e.g., α‐ or β‐diversity). Temporal variation of diversity might be the product of either gradual or precipitous environmental changes including land cover alterations and climate warming (Legendre, 2019). Hence, unraveling the relationships between taxonomic, functional and phylogenetic temporal β‐diversity will improve our understanding on how communities are modified through time under anthropogenic pressures, to further predict future changes and propose effective conservation tools for the potential resilience of ecosystem functioning (Bevilacqua & Terlizzi, 2020; Devictor et al., 2010; Hooper et al., 2005).